Skip to content

Refine Attribute Resolution #97

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
davepeck merged 23 commits into from
Jan 9, 2026
Merged

Refine Attribute Resolution #97

davepeck merged 23 commits into from
Jan 9, 2026

Conversation

@ianjosephwilson
Copy link
Contributor

@ianjosephwilson ianjosephwilson commented Dec 26, 2025 •
edited
Loading

Refine attribute resolution to allow for merging of class and style attrs and smooth out a few edge cases.

Summary

  • Accumulate class values together internally as a dict[str, bool] and output to a str() or None to omit when empty.
  • Accumulate style values together internally as a dict[str, str|None] and output to a str() or None to omit when empty.
  • Expand aria and data as before but do not omit None or False during intermediary expansion.

Major Changes

  • class changes
    • Merge from left to right except when not forced to merge, ie. pass through singleton literal class.
    • If merging is active then always split str() values into a list of classes.
    • Don't allow sequences within sequences, ie. ['btn', ['btn-primary']]
    • Don't allow dicts within sequences, ie. ['btn', {'btn-primary': True}]
    • Retain False values from dict form to allow classes previously merged to be "toggled" off.
    • Move class handling into processor.py, drop classnames() helper.
  • style changes
    • Merge from left to right except when not forced to merge, ie. pass through singleton literal style.
    • If merging is active then always extract str() values into property name / value dictionary.
    • Retain properties with None value from dict form to allow styles previously merged to be "toggled" off.
  • aria / data changes
    • Do not omit attributes with values of None or False right away so that they may override prior values.
    • Ignore if None to allow for def Comp(img_data=None) type pass-through optional arguments

Types

Just for a quick understanding of how this could look these would be for the non-literal values:

type ClassValueItem = None | str
type ClassValue = ClassValueItem | dict[str, bool] | Sequence[ClassValueItem]
type StyleValue = None | str | dict[str, str | None]
type AriaValue = None | dict[str, object]
type DataValue = None | dict[str, object]

# For example
class_list: ClassValue = ["btn"]
class_dict: ClassValue = {"btn-primary": True}
button_t = t'<button class="theme-holiday" class={class_list} class={class_dict}>Save</button>'
assert str(html(button_t)) == '<button class="theme-holiday btn btn-primary">Save</button>'

@ianjosephwilson
Wrap up low level placeholder handling into PlaceholderConfig helper …
b2a0ea0 
…class.
@ianjosephwilson
Draft version of special class handling with merging and distinct seq…
ca44b4a 
…uence vs dict typing.
@ianjosephwilson
Use True instead of None since we only init from literal.
b3ca458
@ianjosephwilson
Add style equivalent to class merging.
decae08
@ianjosephwilson
Remove old style processing code.
3acbb4b
@ianjosephwilson
Add mixed class types back in but without sequence nesting.
f3290a2
@ianjosephwilson
Allow attributes created via aria dict to clear literal attributes.
b83ed6d
@ianjosephwilson
Allow attributes created via data dict to clear literal attributes.
8616767
@ianjosephwilson
Move special class and style handling state and logic into helper cla…
662fe24 
…sses.
@ianjosephwilson
Expand style testing.
119fffb
@ianjosephwilson
Sp.
e0b8c8e
@ianjosephwilson
Try to uniform design with prior processor.
bc97947
@ianjosephwilson
Make things *right.
87bea0e
@ianjosephwilson
Rename to expander to accommodate allocator, pull up into main resolu…
93e8d40 
…tion.
@ianjosephwilson
Re-format.
50ef974
@ianjosephwilson
Remove excess docstrings.
df2708e
@ianjosephwilson
Adjust docs, drop classnames helper.
c495ab8
@ianjosephwilson ianjosephwilson mentioned this pull request Dec 26, 2025
Closed
@pauleveritt
Copy link
Contributor

pauleveritt commented Dec 26, 2025

Good timing, I'm talking to @koxudaxi about how to extract the rules into a spec for an LSP. If you feel confident these changes are final, I will jot them down.

@ianjosephwilson
Copy link
Contributor Author

ianjosephwilson commented Dec 26, 2025

Not quite confident yet, I think we'd need to merge them and then try them out a bit and see if they are annoying to use. I'm not sure what format would be best overall since there are multiple things all intermingled together to get to the end result.

@pauleveritt
Copy link
Contributor

pauleveritt commented Jan 2, 2026

@ianjosephwilson Looks like a policy I should get into the specification stuff I was doing for @koxudaxi ... if you can jot some thoughts down in a reply about styles or other attribute edge cases, I'll capture them.

@pauleveritt pauleveritt closed this Jan 2, 2026
@pauleveritt pauleveritt reopened this Jan 2, 2026
@pauleveritt
Copy link
Contributor

pauleveritt commented Jan 2, 2026

Apologies, clicked as my cursor travelled over the Close with comment button

@ianjosephwilson
Copy link
Contributor Author

ianjosephwilson commented Jan 2, 2026

@pauleveritt Do you mean the rules for how to validate a t-string used by tdom ?

@pauleveritt
Copy link
Contributor

pauleveritt commented Jan 2, 2026

@ianjosephwilson Sorry I though I mentioned in a discussion. @koxudaxi is reviving his PyCharm extension which teaches the IDE the tdom syntax. In particular, autocomplete and validation of props.

As part of that, I did some stuff to decipher the tdom "spec" into a JSON file he could use as the rules.

@ianjosephwilson ianjosephwilson marked this pull request as ready for review January 5, 2026 20:20
@ianjosephwilson
Copy link
Contributor Author

ianjosephwilson commented Jan 5, 2026

@davepeck Later this week when you get a chance I think this is a strong contender for merging but look it over. I'm going to put another note in the related issue. I just had it drafted because it was stacked on top of the other small PR you merged earlier.

@davepeck
Copy link
Contributor

davepeck commented Jan 8, 2026 •
edited
Loading

@ianjosephwilson Thanks so much for this PR and for your patience while I, uh, put my feet up on a tropical beach for a couple weeks. :-)

After reading this over in detail, I think it's a great stake in the ground and I'm happy to merge it and push an update to PyPi. My instinct is that we'll need to use it for a while to see how we feel about the behavior. I'm happy to make adjustments going forward, including large breaking changes (still pre-alpha!), as we learn more.

Does this sound good to you? If you're feeling good about the PR, I'll go ahead and merge.

PS: good riddance to classnames(). 😅

@ianjosephwilson
Copy link
Contributor Author

ianjosephwilson commented Jan 9, 2026

Sorry I keep stacking commits on. This is a little bit more restrictive, as mentioned in #89 , but let's try this out and if we keep hitting cases where we need the other forms we can add them back in. A good overview is in def test_interpolated_class_attribute().

@davepeck
Copy link
Contributor

davepeck commented Jan 9, 2026

@ianjosephwilson Those changes look good. TBH I'm somewhat skeptical that toggling will be easy for the typical dev to reason about in practice, but let's give it a go. Merging.

@davepeck davepeck merged commit 0c4d7ab into t-strings:main Jan 9, 2026
1 check passed
@ianjosephwilson ianjosephwilson deleted the ian/refine_attr_resolution branch January 9, 2026 18:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@ianjosephwilson @pauleveritt @davepeck